Covariance measures how two variables vary together: positive if they tend to increase together, negative if one increases when the other decreases. Its value depends on the units and is hard to interpret on its own. Correlation is its normalized version, between -1 and 1.
Example
df['x'].cov(df['y'])